opencv - 3D Reconstruction upto real scale -
i working on project detect 3d location of object. have 2 cameras set @ 2 corners of room , have obtained fundamental matrix between them. these cameras internally calibrated. images 2592 x 1944
k = [1228 0 3267 0 1221 538 0 0 1 ]
f = [-1.098e-7 3.50715e-7 -0.000313 2.312e-7 2.72256e-7 4.629e-5 0.000234 -0.00129250 1 ]
now, how proceed given 3d point in space, should able points on image correspond same object in room. if can obtain right projection matrices (with correct scale) can use them later inputs opencv's traingulatepoints function obtain location of object.
i have been stuck @ since long time. so, please me.
thanks.
sorry not coming in size of comment.. @user2167617 reply comment. pretty much. few pointers, though: singular values should (s,s,0), (1.3, 1.05, 0) pretty guess. r: technically, right, however, ignoring signs. might rotation matrix not satisfy constraint deteminant(r) = 1 instead -1. might want multiply -1 in case. generally, if run problems approach, try determine essential matrix using 5 point algorithm (implemented newest version of opencv, have build yourself). scale indeed impossible obtain these informations. however, it's scale. if define example distance between cameras being 1 unit, measured in unit.
Comments
Post a Comment